SIO 221C Data Analysis Laboratory

Fall 2013

Professor: Sarah Gille

SIO Office: Nierenberg Hall 348
Telephone: 822-4425
e-mail: sgille at ucsd.edu

Meetings: Tuesday and Thursday: 12:30-1:50, OAR Conference Room

Course Requirements: Complete 3 projects from list below (or new ones to be added; the final project may be arranged with instructor to be tailored to research). Each project is scheduled to take 3 weeks. Students report on their progress at each class meeting and discuss problems and possible solutions with the group. Written reports are submitted at the end of each project. The course is taken S/U.

Course evaluations (Nov 25-Dec 9): You should have received an e-mail from UCSD Academic Affairs explaining how to evaluate this course. Please take the time to submit an evaluation, since it provides valuable feedback.

Syllabus

Resources:

Reference books
Selected papers
Software possibilities

Schedule:

Organization: September 26

Project 1 (spectral methods): October 1, October 3, October 15, October 17, October 22, October 24; written reports due October 24)

Project 2 (EOFs or objective mapping): October 29, October 31, November 5, November 7, November 12, November 14; written reports due November 14)

Project 3 (open choice): November 19, November 21, November 26, December 3, December 5, December 10; written reports due December 10)

No class due to instructor travel (October 8, October 10); Thanksgiving (November 28)

Submit papers to turnitin.com using these instructions. I'd also be grateful for a hard copy.

Projects

Time series and spectra
- Salinity Spiking (Rudnick)
- Filtering and Assessing Resolution (Gille)
- Wind-Driven Currents (Rudnick)
- Complex Demodulation (Rudnick)
- Wind-Driven Geostrophic Currents (Gille)
- Wavelets (Rudnick)
Empirical Orthogonal Functions
- Empirical Orthogonal Functions (Rudnick)
- EOFs with Missing Data (Rudnick)
Objective Mapping
- Objective Mapping (Rudnick)
- Objective Mapping (with Anisotropic Decorrelation Functions) (Gille)
- Geostrophic Velocity (Rudnick)
Other
- Box Inverse (Rudnick)
- Probability Density Functions (Gille)

Data

Tools

NetCDF for Matlab. (And some tips for using netcdf.)
grabcdf.m
Thermodynamic Equation of Seawater - 2010 sea water toolbox, McDougall and Barker (2011). (If you look at older work, you might see reference to the CSIRO Seawater Toolkit.)
cptcmap.m: Function to create Matlab color tables using palettes created for the Generic Mapping Tools (GMT).
stick.m
arrow.m
m_map mapping package for use in Matlab, developed by Rich Pawlowicz (UBC).
Mathworks file exchange for user contributed Matlab scripts (suggested by Gabriela Chavez)
NetCDF Operators (NCO)
Comments on using LaTeX to write manuscripts (plus a short guide on equations from the American Mathematical Society web site).
A short guide to LaTeX
angsep.m computes distance between two geographic points in degrees
significance.m simple routine for computing statistical significance levels for correlation coefficients. For example, for 8 degrees of freedom, the 95% significance level would be "significance(8,.95)", which is 0.693.

Lecture Notes

Error propagation (September 26)
A refresher on spectra (October 1)
Comments on spectra and spectral error bars (October 3)
Coherence (October 15)
Variance preserving spectra (October 15)
Filters (October 15)
Suggestions for clear writing (October 17)
Edge effects in spectra: detrending, windowing and pre-whitening (October 22)
Choosing color maps (October 29)
Popular press readings on statistics (October 29)
Idealized examples for empirical orthogonal functions (October 31)
Details for EOFs: Uncertainties, alternative methods, propagating modes (November 5--not all discussed in class)
Estimating degrees of freedom from autocorrelations (November 12)
Using probability density functions (November 19)

Sarah Gille's Home Page

Daniel L. Rudnick's Home Page